Lateralized Sensorimotor Evoked Potentials during Visuomotor Transformation in Real and Imagined Movements

Abstract

The neural mechanisms underlying motor preparation have attracted much attention, particularly because of the assertion that they are similar to the mechanisms of motor imagery (MI), a technique widely used in motor rehabilitation and brain-computer interfaces (BCIs). Here we clarified the process of visuomotor transformation for the real and imagined movements by analyzing EEG responses that were time locked to the appearance of visual targets and movement onsets. The experimental task required responding to target stimuli with button presses or imagined button presses while ignoring distractors. We examined how different components of movement-related potentials (MRPs) varied depending on the reaction time (RT) and interpreted the findings in terms of the motor noise accumulation hypothesis. Furthermore, we compared MRPs and event-related desynchronization (ERD) for overt motor actions versus motor imagery. For the MRPs, we distinguished lateralized readiness potentials (LRPs) and reafferent potentials (RAPs). While MRPs were similar for the real and imagined movements, imagery-related potentials were not lateralized. The amplitude of the late potentials that developed during motor imagery at the same time RAPs occurred during real movements was correlated with the amplitude of β-ERD. As such they could have represented sensorimotor activation triggered by the imagery. LRPs that occurred during real movements lasted longer for longer RTs, which is consistent with activity accumulation in the motor cortex prior to overt action onset. LRPs occurred for non-target stimuli, as well, but they were small and short lived. We interpret these results in terms of a visuomotor transformation, where information flows from visual to motor areas and results in a movement, a decision not to move and/or a mental image of a movement. The amplitude of the late positive peak that developed during MI was correlated with the amplitude of the β-ERD. Since the latency of this component was consistent with the timing of RAP, we suggest that it is a non-lateralized RAP-like component associated with sensorimotor activation during kinesthetic MI.